Learning from experience

“There is only one thing more painful than learning from experience and that is not learning from experience.” – Archibald MacLeish

This quote aptly describes my objective in writing my most recent article—“Lessons Learned From Amazon”—about my experiences as Amazon.com’s head of data mining and personalization. It was great to review and articulate some of the main learnings I came out of Amazon with. What were some of our best decisions? Worst? What we did we learn? The wisdom I obtained from the experience—the good, the bad, and even the ugly—has been invaluable and truly applies to all etailors no matter the size.

MultiChannel Merchant published the article this past week. I’ve included the full text below.


Lessons Learned From Amazon

Feb 23, 2009 12:39 PM, By David Selinger

As the former head of Amazon.com’s data mining and personalization team, I gained ample wisdom from what the retail giant does right. But I also saw firsthand a few things Amazon didn’t do quite so well, and learned the hard way that even simple mistakes can cost time, sales and brand equity.

As an e-tailer, you are obviously looking for every possible tool, trick and insight to help boost sales across channels—but have you gone back to basics lately? Here are some simple lessons I learned at Amazon that can help you avoid pitfalls and instead drive sales by enhancing the shopping experience.

1. Site performance is a business problem, not a technology problem.
Whether you sell more in your shops or online, as a multichannel retailer, your online channel is extremely important to driving sales, boosting customer loyalty and building your brand.

We quickly learned at Amazon that the “one second rule” is key to boosting sales. If it takes longer than one second for a page to load, that’s too long. Your customer will get frustrated and go somewhere else. And if your site goes down because of network problems at your hosting company – well, that’s just death to your sales for the time that the site is down – and it can do lasting damage to your brand.

My advice is to carefully vet your site hosting provider, continually test your site to make sure performance meets your requirements and demand always-on, 100% up-time from your content delivery network (CDN) and other vendors. Choose vendors that leverage distributed data centers and the power of parallel computing to offer a truly fail-safe infrastructure; your customers will thank you by returning to your site often, a critical component to growing sales.

2. Your data are your greatest asset. I also call this the “brother-in-law” lesson, because while brothers-in-law are often smart and may have good golf tips, their advice about e-commerce, while well intentioned, may not be so well informed.

At Amazon, we learned the hard way that data should always be the greatest source of wisdom. After attempting some cool-sounding product enhancements attained from numerous brothers-in-law without much success, it became clear that our data delivered the smartest insights about e-commerce sales; clicks and conversion don’t lie.

If you are a multichannel retailer, it’s even more critical to craft an e-commerce strategy around careful analysis of your online sales data, because the merchandising, promotional and marketing techniques that work in stores or in catalogs don’t always work online.

You need a sophisticated e-commerce strategy based on rigorous testing and analysis of your customer, traffic and sales data. Invest in your data-warehousing infrastructure. Perform A/B testing on campaigns and pages. Use Web analytics tools to figure out why people are coming to your site and what they hope to find. Use data analytics tools to figure out what people want to buy, and why.

3. Simplicity sells. Your customers don’t care about a jazzy, flashy site; they want the best deals on the best products delivered to them quickly and efficiently. When I say simplicity, I mean presenting the products people want and making it easy for them to click and buy.

Amazon.com got two things right in the simplicity department: providing product recommendations to customers via a single line of text at the top of the page, reading “Welcome John Doe! We have product recommendations for you in books, music and more;” and rolling out its famous “one click” buying, which allows users to check out in a flash. This simple messaging of personalized recommendations drove a significant boost in site-wide sales.

But Amazon also got simplicity wrong in some instances. For example, the merchant tried to make some shopping pages less “busy” and give them a Google-like aesthetic. The pages looked cool, but sales and conversion dropped.

It seems that Amazon customers liked the original look and feel of the Amazon site, offering dozens of products and choices—as long as it was “simple” to find products they wanted and easily buy them. Of course, some more complex projects can yield great value, but be sure you’ve got your basics down first, and test the heck out of anything you change!

4. Recommend—and recommend right. One of the things Amazon got right was personalized product recommendations. The retailer was an early leader in the field of “product discovery,” and its big push into product recommendations boosted site-wide profits by 10%-20%.

As BusinessWeek journalist Sarah Lacey put it in an article last December: “Mediocre discovery doesn’t do anything but annoy people. But done right, discovery delights customers, fosters loyalty and gets people to spend more. To date, Amazon has done this better than anyone.”

These results didn’t come freely or easily—lots of testing and lots of tuning—but the results speak for themselves. High-quality recommendations are fundamentally better than irrelevant ones.

Further, recommendations don’t stop with the product page. Product recommendations should be integrated throughout the shopping cycle – from homepage through to shopping cart to checkout, and even in marketing campaigns.

Initially, Amazon sent out blanket e-mail messages such as “10% off today,” but it didn’t include personalized product recommendations. By adding recommendations framed with appropriate context such as “people who purchased this item also purchased this item” in e-mails, Amazon dramatically boosted e-mail-based sales.

Brick-and-mortar retailers should add recommendations to in-store kiosks, and have customer service operators make product recommendations, too.

Amazon.com is a trailblazer in e-commerce and one of the most successful retailers of all time. But it’s not immune to making mistakes, and we can all learn from these lessons to deliver a better e-commerce experience to our customers.

Solidify your site infrastructure, simplify your product offerings and checkout systems, mine your customer data and integrate high-quality recommendations throughout the shopping process. Ace these four things and you’re sure to boost sales by delighting your customers.

Share :

This post was written by David Selinger

ABOUT David Selinger
David is CEO and founder of RichRelevance. He first garnered international recognition as an expert in the field of eCommerce data analytics and personalization with his groundbreaking work leading the research and development arm of Amazon’s Data Mining and Personalization team. In that role, David increased Amazon’s annual profit by over $50 million (25% of US profit, 2003) setting the industry standard for recommendation services. To view David's full profile, click here.
Related Posts

Leave Your Comment